I was asked by the Notices of the AMS to review the book “More Sex is Safe Sex: The Unconventional Wisdom of Econmics” by Steven E. Landsburg. My review entitled “Economics and Common Sense”, will appeared in the June/July issue of the Notices and you can find it here. (And in the August issue of the AMS Notices, there is will be a book review by Olle Häggström on John Allen Paulos’ new book: “Irreligion”.)

In his book, Steven E. Landsburg uses the “weapons of evidence and logic, especially the logic of economics” to draw surprising insights which run against common sense. “If common sense tells you otherwise,” says Landsburg, “remember that common sense also tells you that the Earth is flat”.

I will include a few little sectionettes from the review here in this post. Some of the issues raised in this book are related to many discussions and debates we had over the years at the Center for the Study of Rationality of the Hebrew University of Jerusalem. Questions regarding “efficiency,” “subsedies,” “monopolies,” “labor union,” “differential salaries,” “law and economics,” “rationality and the judicial system,” and various other related topics were amply discussed at the Center, and some of these topics and discussions are related to issues raised in Landsburg’s book.

**The Armchair Economist**

Sometime after writing my review, I saw in the Yale book store, along with several copies of Landsburg’s new book, a copy of his older book from the early 90s, “The Armchair Economist”. I bought it and read some of the chapters. So before quoting excerpts from my book review on the new book let me first talk a little bit about the older book. Several chapters are devoted to describing some classic teachings of economics, such as general equilibrium theory, and they are very good. I liked Landsburg’s explanation of the notion of “efficiency” – it is the best popular explanation of “efficiency” I read or listened to.

**The Peltzman Effect**

*Armchair Economist* starts with an interesting finding from the 1970s by Sam Peltzman, a U. of Chicago economist, asserting that safety seat belts have led to an increase in the number of car accidents, and that the net effect on death from car-accidents, which is reduced by the direct effect of seat belts and other safety measures, but which at the same time is increased by the indirect effect of drivers taking more risks because of these safety measures, is close to zero. Peltzman’s study also asserts that seat belts have raised pedestrian death toll caused by accidents.

Landsburg begins the story with the influential 1965 book by Ralph Nader “Unsafe at Any Speed” that led to wide range automobile safety legistlation, including obligatory seat belts. According to Landsburg, any economist could have predicted that these measures would lead to an increase in car accidents, because safe cars give an incentive to drive less carefully. (For those who find it hard to believe that people drive less carefully because cars are safer, Landsburg suggests considering the proposition that people drive more carefully if their cars are more dangerous.) And indeed this was supported, according to Landsburg, by Peltzman’s study.

I must admit that I was (and still am) quite skeptical about Peltzman’s effect and particularly skeptical about the quantitative aspects. Here is an analogy. In the last two decades medical treatement for lung cancer has led to longer survival of lung cancer patients. Is it reasonable to assume that this medical development gives a noticable incentive for smoking more? Would you find credible a claim that the effects of better treatement for lung cancer and more smoking, followed by better treatment for lung cancer, cancel each other?

Peltzman’s effect appears to be quite well-known and it is discussed and quoted even in standard economics textbooks. I was wondering what more recent studies on the effect of safety measures in cars over the last three decades tell us, and whether they support Peltzman’s (quantitative and qualitative) conclusions. Is Peltzman’s effect real and substantial? Was it critically studied and was it sufficiently replicated? Of course, a lot of material can be found on the Internet. Let’s return to this topic sometime. Two interesting links: A paper by Alma Cohen and Einav Liran, and a recent lecture by Peltzman. (Update: links fixed, May 31.)

One other nice thing in the older book is Landsburg’s reference to people who object to new oil drilling as “mineral rights activists”- that’s quite funny!

### Three remaks:

1) There is a great line by **Malka Heller** (Yuval Peres’ grandmother), which is relevant to many fantastic claims that run against common sense: Heller used to say:

### Why should I be surprised if I can simply disbelieve.

2) The new book by Landsburg gives more weight to remote and isoteric applications of “the logic of economics,” while the older book gives more weight to the classical models and insights. Does this change represent a shift of interest of economists in general?

3) In reality, the “logic of economics” is often required to confront not “the logic of mathematicians,” but rather the “logic of law” (and sometimes, the logic of politics). Many of the arguments from the economics side which seem counterintuitive (and incorrect), should be examined in view of opposing arguments (which also seem to contrast with common sense) that derive from legal considerations and principles.

**************************************************

**Here are three little sections from my book review:**

In his book, Steven E. Landsburg uses the “weapons of evidence and logic, especially the logic of economics” to draw surprising insights which run against common sense. “If common sense tells you otherwise,” says Landsburg, “remember that common sense also tells you that the Earth is flat”.

**Is more sex safer? **Let us start with the first example** **that gives the book its name. Common sense tells us that promiscuity spreads AIDS. Landsburg, relying on a paper by Harvard economist Michael Kremer, argues otherwise. Before presenting Landsburg’s claim let’s make sure we know what he does not claim. Landsburg agrees that for *every* individual in the society, having more sex is less safe. (Landsburg also agrees that practicing safe sex is safer than not taking safety measures.) Yet he argues that if sexual conservatives relax their standards, sex will be overall safer. “Michael Kremer estimates that the spread of AIDS in England could possibly be retarded if everyone with fewer than 2.25 partners per year were to take additional partners more frequently.” We can imagine the mathematical stochastic model behind this claim: men and women are presented by vertices in a bipartite graph which is used to describe the spread of the epidemic. A sexual relation is described by an edge and the main observation will be based on the epidemic being spread more rapidly if the variance in the degrees of vertices is larger. (There is not much about sex, neither in this model nor in Landsburg’s book as a whole.) Is this argument convincing? Does it represent a solid contribution of economics theory (and even some mathematics) to the area of Medicine? Should it be translated to practical social recommendations? I was not convinced. It seems to me like a case when you base an analysis on second-order effects and neglect the first order ones. But I will let you read the book, or better yet, Kremer’s original paper, and make your own judgment. To be sure, this is provocative and quite interesting.

**The role of reason and mathematics. **The same chapter 12 ends with a lovely defense of the application of pure reason and mathematical modeling to social issues. I heartily recommend reading it. Landsburg says that “Resistance to logic frequently reveals itself as animosity towards mathematics” and mentions readers that claimed that “no mathematically expressible argument can ever be relevant to a moral dilemma.” Landsburg disagrees and I agree with him. The major role of mathematics and statistics in economics and other social sciences is a 20^{th} century development. Economist Herbert Scarf described to me the unique role of the Cowles foundation at Yale University in bringing about the dominance of mathematical and statistical methods in economics. (He “complained” though that the success of this revolution has made *Cowles *a less unique place than it used to be.) But mathematicians can be just as skeptical about applications of mathematics to social sciences as those with animosity towards mathematics. One difficulty in the interpretation of mathematical modeling and results is that often they run much beyond the scope of the original mathematical setting. As mathematician Wolfgang Dahmen often argues, the most important thing to remember regarding the application of mathematics is the sentence: “If if … then then”. The conclusions of a mathematical theorem go only as far as its conditions allow. Right? Well, in most areas where mathematics is applied the interpretations go well beyond what mathematics allows. This accounts for the many successes in applications of mathematics and mathematical formalism, and for quite a few failures, as well. Mathematical formalism is useful not only to support good ideas but also to shoot down as meaningless or incorrect bad ideas.

**A little riddle:** And my review includes a little riddle inspired by Landsburg bold argument relating the logic of economics and the flat earth intuition. Can you solve it?

You sail 300 kilometers from point A to B and then in a perpendicular direction 400 kilometers from point B to C what is the sailing distance from A to C. Choose the best answer. a) 500 kilometers b) 499.99 kilometers c) 490 kilometers d) 450 kilometers

**************************************************

LucasUsing the spherical pythagorean theorem Cos(c/r)=Cos(a/r)*Cos(b/r), and the radius of the earth being 6378.1km, my calculator gives 499.82, so none of the above. Of course, the intuition is that the third side should be shorter than the pythagorean theorem predicts, since we are living on a roughly spherical object. Since the earth is very close to being flat, the best answer is (b).

GilDear Lucas, I think that for this particular “sailing distance” problem, the “flat” approximation is probably better for every practical purposes. It is certainly better, here, if you cannot estimate, even roughly, the “correction” needed, based on the earth being round, which is the case for answers b), c) and d). Therefore, in my opinion, a) is the “best” answer.

JKFor the riddle, doesn’t it actually depend where on the earth you start from? In particular, one can imagine setting the numbers and starting point close enough to the North pole such that going West by x and then North by y would get you right back to your starting point…

JKLooking at the reference you give, I see there is the requirement that the lines of the triangle correspond to arcs of a circle centered at the center of the sphere. This would not be the case in the example I suggested — but I think this requirement is also unclear in your description of the problem.

LucasHere’s an interesting link about a supposed natural experiment testing the name of the book: http://www.marginalrevolution.com/marginalrevolution/2008/06/more-sex-is-saf.html

GilVery interesting. Thanks, Lucas.

Yechezkel ZilberSeat belts and lung cancer example.

First, I assume that people value for staying alive is not absolute. Fact: they drive when they can use safer ways. (I guess it is in an abstract sense maginal. While in practice it does not follow a mathematical line, but the mathematical line is a theoretical approximation, as in many other cases).

Lung cancer is not less aversive because of treatment. Staying alive and suffering is not much better than dying.

And why do you think that IF lung cancer becomes less aversive it will not have an effect on smoking decisions?

There is always a complex route from effect to decisions. Smoking also implies very long term effects, while driving is probabilistic. And very many people have accidents. I had an accident. Car was totally lost, but I only had blood in my nose. Needless to say that the effect of this acident of my future driving was small. I am sure that breaking a leg would have had much more effect.

Pingback: Plans and Updates « Combinatorics and more

Derek JonesFor data (ie, evidence backing up the economists seat belt claim) and

discussion on the car safety issues see http://john-adams.co.uk/

With respect to lung cancer, modern medicine is detecting cancer earlier

and people are living longer after detection. However, they are not living

longer in absolute terms, it is just that the starting point on a relative

scale has moved back in time.

Gil KalaiDear Derek,

Thanks for letting me know the difficulty you had in posting the comment. Indeed the blog by John Adams is quite interesting; We will have to come back to the “Peltzman effect” regarding seat belts regulation sometime. (There are different views among economists on this matter.)

BenI’d say none of the answers to your riddle are satisfactory. If I sail 300 km one way then 400 km perpendicular to my original path, it’s probably because I’m sailing around a peninsula in which case the shortest sailing distance from A to B is probably slightly less than 700 km.